Does school average achievement explain the effect of socioeconomic status on math and reading interest? A test of the Information Distortion Model